A facile, sensitive and highly specific HPLC method for assaying 1-(2-chloroethyl)-3-sarcosinamide-1-nitrosourea (SarCNU) in plasma has been developed. The drug was efficiently isolated from plasma by extraction with tert-butyl methyl ether. A structurally related compound with similar physicochemical properties served as the internal standard (IS). Following evaporation of the organic solvent, the extract was reconstituted with 0.05 M ammonium acetate buffer, pH 5.0, and loaded onto a 4 microm Nova-Pak C18 column (15 cm x 3.9 mm), which was preceded by a 7 microm Brownlee RP-18 precolumn (1.5 cm x 3.2 mm). Chromatography was performed at ambient temperature using a mobile phase of methanol-0.1 M ammonium formate buffer, pH 3.7 (25:75, v/v). UV absorbance of the effluent was monitored at 240 nm. A flow-rate of 1.0 ml/min was used for analyzing mouse and dog plasma extracts. An automatic switching valve was employed to allow the precolumn to be flushed 1.5 min into the run, without interrupting flow of the mobile phase to the analytical column, thereby preventing the apparent buildup of extractable, strongly retained, UV-absorbing components present in mouse and dog plasma. Operating in this manner, more than 100 samples could be analyzed during a day using a refrigerated autosampler for overnight injection. Mean values of the tR for the drug and IS, determined from chromatograms of a 1 microg/ml mouse plasma standard during an 8 week period, were 3.99 plus minus 0.08 and 6.12 plus minus 0.17 min, respectively (S.D., n = 27). The method was readily adapted to the determination of SarCNU in human plasma by simply decreasing the eluent flow rate to 0.6 ml/min, which afforded tR (mean plus minus S.D., n = 10) of 5.78 plus minus 0.03 and 9.07 plus minus 0.06 min for SarCNU and the IS, respectively. Furthermore, the switching valve was not necessary for the analysis of human plasma samples. With a 50 microl sample volume, the lowest concentration of SarCNU included in the plasma standard curves, 0.10 microg/ml, was quantified with a 7.8% R.S.D. (n = 27) over a 2 month period. Plasma standards having concentrations of 0.26 to 5.1 microg/ml exhibited R.S.D. values ranging from 1.3 to 4.7%. Thermospray-ionization MS detection was used to definitively establish the specificity of the method. The sensitivity of the assay was shown by application to be more than adequate for characterizing the plasma pharmacokinetics of SarCNU in mice.